Validation of a Metroticket Approach to prognosticate overall survival (OS) after Surgical Resection for Hepatocellular Carcinoma (HCC): a multi-center study

Abstract

Introduction: Surgical resection (SR) for HCC is frequently carried out for intermediate-stage HCC outside of the Milan criteria. Recently better loco-regional therapies (e.g. Y90, SBRT) and systemic therapies (e.g. nivolumab) have demonstrated high efficacy in intermediate-stage HCC. Thus surgeons need to predict at which point surgical resection in “out of criteria” patients has better OS over other therapies. A Metroticket approach to prognosis in HCC has previously been validated in transplantation. We aimed to validate this with SR. Method: A retrospective cohort analysis of SR for HCC between 1st January 2000 to 30th June 2016, at the National Cancer Centre Singapore/Singapore General Hospital, was performed. Inclusion criteria were: pathological confirmation of HCC, > 5-year follow-up or death within 5 years and complete data. Following the method of the “Metroticket 2.0 Model” (Mazzaferro et al., 2018), we produced smoothed bivariate contour plots of 5-year HCC-specific OS probabilities based on the pre-operative (1) radiologic Tumor Burden Index (TBI = size of the largest tumor nodule + the number of nodules), and (2) AFP concentration in ng/ml. Bland-Altman (B-A) analysis assessed agreement between TBI calculated from pre-operative imaging versus post-surgical pathological assessment. The model was validated against an independent cohort of 200 patients with HCC resected at the Duke University Medical Center, USA. Result: Of 1121 patients with HCC resected, 601 patients fulfilled inclusion criteria. Iso-survival curves from bivariate contour plots of 5-year HCC specific survival probabilities (p < 0.05) showed good validity (p < 0.05). TBI calculated from Imaging – Histology showed d = 0.13 with sd = 1.76, (95% limits of -1.63 and 1.89). A 5-year OS of at least 50% was found for patients with a TBI of ≤6 and/or AFP of <15 ng/mL. This was confirmed by the validation cohort. Conclusion: A Metroticket approach allows robust prognostication of OS with SR in HCC.